Skip to content

Difficulty Curves in Arcade Games

How arcade games escalate, why the classic curve was shaped by coin-operated economics, and what a difficulty curve has to do to stay fair while getting harder.

Reign Creative Team4 min read

Key takeaways

  • Classic arcade difficulty curves were shaped by coin-operated economics: a run had to end reasonably quickly for the machine to earn.
  • The four levers a designer has are quantity, speed, pattern complexity and margin for error, and they feel very different to a player.
  • Difficulty feels fair when the player can identify what killed them and what they should have done differently.
  • Increasing quantity is the cheapest escalation and the least interesting; increasing pattern complexity is the most interesting and the most work.
  • Endless modes have to escalate without a ceiling, which usually means combining levers rather than pushing any one to an extreme.

Arcade games are hard because of how they made money. A coin-operated machine earns per play, so a run has to end within a reasonable time. Difficulty escalation was the mechanism, and it shaped a genre convention that outlived the economics that produced it.

Understanding that explains both why the genre feels the way it does and what a modern designer is actually choosing when they set a curve.

The four levers

A difficulty curve is built from four variables. They can produce the same measured difficulty and feel completely different.

Lever What it changes How it feels
Quantity Number of enemies or hazards Overwhelming; pressure
Speed How fast things move and how fast you must react Frantic; reflex-testing
Pattern complexity How intricate the behaviour is Demanding; puzzle-like
Margin for error How much punishment a mistake carries Tense; high-stakes

Quantity is the cheapest escalation. Doubling the enemies is trivial to implement and produces a real difficulty increase — and it is the least interesting, because it asks nothing new of the player.

Speed is nearly as cheap and hits a ceiling fast. Beyond a point, faster stops being harder and becomes impossible to react to at all, which is a different thing.

Pattern complexity is the interesting lever and the expensive one. New behaviour is content, and it asks the player to learn rather than merely to execute faster.

Margin for error is the sharpest. Reducing what a mistake costs you changes the emotional register of the whole game — the same content played with one life instead of three is a different experience.

What makes an increase feel fair

A death feels fair when the player can identify what killed them and what they should have done instead.

That is the whole criterion, and it is about legibility rather than about difficulty level. Two consequences:

A brutal but legible death is acceptable. Players will retry something they understand almost indefinitely.

An easy but illegible death is not. A death from an off-screen hazard, or from a mechanic the game never explained, produces frustration out of proportion to its difficulty.

This is why telegraphing matters so much. An attack that signals itself before landing can be far more dangerous than one that does not, because the player has a decision to make. Boss pattern recognition covers how that legibility works in practice.

Escalation within a run

Most arcade games escalate inside a single run rather than between levels. Two common structures:

Wave-based. Discrete groups arrive, each harder than the last. Escalation happens at wave boundaries, which gives the player a rhythm and a moment to breathe. Wave survival game design covers this structure in detail.

Continuous. Difficulty rises smoothly with time or score. Fewer natural pauses, more sustained pressure.

Wave-based is easier to tune, because the designer controls exactly what arrives when. Continuous escalation feels more relentless and is harder to keep fair, because there is no checkpoint at which the player's state is known.

The player's counter-curve

The reason a fixed escalation works at all is that the player improves during the run — and across runs.

Two kinds of improvement:

Within a run: the player's build gets stronger, if the game has progression. Mid-run upgrades are the counter-curve.

Across runs: the player learns patterns. This is why a game that felt impossible on the first attempt becomes tractable on the tenth without anything changing.

A well-tuned curve rises slightly faster than the within-run counter-curve, so pressure increases, but slowly enough that the across-run learning keeps up.

The roguelike tradition made the across-run learning explicit — the Berlin Interpretation is an attempt to define the genre whose core loop depends on it — and what makes a roguelike covers where those definitions hold and where they do not.

Dynamic difficulty adjustment

Some games adjust difficulty in response to how you are doing: easing off after repeated deaths, tightening when you are dominating.

It smooths the experience, and it carries a cost. If players notice it, success feels unearned. A player who suspects the game let them win has lost the thing that made winning worth doing.

The usual compromise is subtlety — adjustments small enough to be invisible — which limits how much they can do.

Endless modes

An endless mode has to escalate without a ceiling, which rules out pushing any single lever to an extreme. Push speed alone and it becomes unreactable; push quantity alone and it becomes visual noise.

The workable approach is combining levers gradually, so that the total pressure rises while no single dimension becomes absurd. An endless mode is really a statement that the only limit is how long your positioning holds — which is the same open-ended score contract the earliest arcade games made.

Space Invaders, documented by the Strong National Museum of Play, is the origin point of that contract, and the history of shoot 'em up games traces where it went from there.

Where our games fit

Space Shooter - Galaxy Arcade uses readable enemy patterns and four distinct abilities — main weapons, a secondary attack, a charged Ultimate and a Shield — so escalation is met with decisions rather than only with faster reactions. Its Endless Mode removes the finish line entirely.

Zombie Survival: Last Survivor escalates on the quantity lever with movement as the only input, and counters it with mid-run level-ups that hand you a choice of upgrade on the spot.

Both are free to download on Google Play, supported by ads, with optional in-app purchases and Everyone content ratings.

Score attack strategy covers playing these curves well rather than designing them, and everything we make here is under action and arcade games.

Good to know

Frequently asked questions

Why were classic arcade games so hard?
Because of the business model. A coin-operated machine earns per play, so a run had to end within a reasonable time. Difficulty escalation was the mechanism, and it produced a curve much steeper than a home game needs.
What levers does a designer have?
Broadly four: how many enemies or hazards there are, how fast they move, how complex their patterns are, and how much margin for error the player has. Each produces a different feel even at the same measured difficulty.
What makes difficulty feel fair?
Being able to identify what killed you and what you should have done instead. A death you cannot explain feels arbitrary regardless of whether it was, and a death you understand feels fair even when it was brutal.
What is dynamic difficulty adjustment?
A system that changes difficulty in response to how the player is doing. It smooths the experience and has a cost: if players notice it, it can make success feel unearned.
How do endless modes work?
They have to escalate without an end point, which means combining levers gradually rather than pushing any single one to an extreme. Pushing one alone eventually produces something unplayable rather than hard.

Sources

  1. Space Invaders (opens in a new tab)The Strong National Museum of Play — accessed
  2. Berlin Interpretation (opens in a new tab)RogueBasin — accessed

Apps mentioned in this article